And students know what answer the professor wants to hear, and he is giving them bonus points on their grade for participating in the experiment. Not to mention that a student who participates in the bonus-point offer is likely a student who is doing badly in the class and would appreciate the assistance from the points and from removing the distractions, as they likely are the same students that think the professor knows what he's talking about, otherwise why would they volunteer for the experiment at all?

In that situation, it would not be difficult to cherry pick some supporting opinions. Remember, this is not exactly a controlled experiment.